Mesh Extender assembly

Following our crowd-funding campaign we have funds to build a number more mesh extenders, and get them out to some early adopters as well as to NZ Red Cross.  So the last few days have been a bit of a production line while we get the hardware assembled.

The worst part of the assembly is getting the lids of the MR3020 routers, which left me with a broken pocket knife and sore fingers for a couple of days.  Basically the lids are cemented to the main body of the case, and you have to break each point of cemented bond, without breaking the lid.  Some are much easier than others, depending, presumably, on the amount of cement used.  We had one lid cracked out of the ten units made.

Then it was soldering the radios and router PCBs together:

The cases are down with the engineering workshop services to get the holes drilled for the radio connectors, and some of the internal ribbing ground off so that the radio can sit flush with the inside of the case.

Otherwise, it has been teaching, and sorting out some of the local tax arrangements for the contributions from Australians in our crowd funding campaign.  If you put an Australian shipping address and requested a perk, then you will get a GST-inclusive tax invoice.  We won't be out of pocket with GST, because we budgeted for it, and also the GST on all the hardware will more than make up for the GST we have to pay on Australian contributions.

Our focus is now on fixing a couple of late issues with the mesh extender firmware so that it works at least for meshms over UHF radio.  Optimising the throughput for larger files will happen overtime, and be delivered as over-the-mesh updates.

Avoiding jet-lag completely, and other body-clock hacks.

Jet-lag is not fun.  Nobody likes jet-lag.

But for me, the situation is more acute.

First, I live in Australia, so travelling almost anywhere involves substantial time-shifts. I still remember my surprise when going to a conference in China and realising that: (a) I wasn't jet-lagged; and (b) everyone else was instead. Such is the effect of living in Australia.

Second, I have a young family at home, and also teach at a University.  This means that most of my trips have to be kept to the absolute minimum duration.  Take my two most recent trips:

Trip 1: Adelaide -> Washington DC -> Amsterdam -> Adelaide in 9 days, including a complete 24-hour time shift, and heading East all the way, which is reputed to be the worst direction for jet-lag (a subject I'll comment on later).

Trip 2: Adelaide -> London -> Adelaide in 95 hours. That's right, I wasn't even gone for four whole days. But I had to be fully in the London timezone immediately for meetings two hours after landing, and then on the one whole day I was there because I was presenting a grand-final pitch at the Global Security Challenge.  I then had to be back in Adelaide time straight away on returning, because I was rostered on at church, and then had a meeting that afternoon.

If I had tried these trips three years ago, it would have been jet-lag city.  Fortunately since then I have learned progressively more about an excellent and (for me at least) easy way to beat jet-lag before it even happens.

This is the fasting method of which there are lots of descriptions online, including this one.  But I want to add to the record, first by confirming that for me it works, and second, outlining my simplified approach, and some comments about one of the causes of killer week-long jet-lag that has been revealed by the discovery of the ability of the hunger clock to override the circadian body clock.  But first the method:

The method

1. Before departing, set watch to destination timezone.
2. Don't eat after 16h00 in destination time zone.
3. Try to sleep as much as you can on the plane so that you aren't sleep deprived on arrival (which is distinct from jet-lag, but no more fun). Pack an eye-mask or use the one they supply on board.  If you can manage it, snooze all the way on a 24 hour flight, with the odd movie etc to keep you occupied if you like.  The sleep part isn't important for the jet-lag, however, just your sleep deprivation, for reasons I will soon explain.
4. Do eat at 08h00 or similar sensible breakfast time in destination time zone.

When I say don't eat, I mean don't eat. Nothing. No fruit. No Juice. No oysters, roast turkey, custard pies or anything. Only water.

Why you can't even eat a little bit during the fast

The reason it is really, really, important that you don't eat during that 16 hour period is that your body needs to switch to "hungry mode", and get the idea that during that time there is no food coming. If you eat anything, no matter how small, during the fasting period, it is basically game over, go to jail, do not collect $200. Try again tomorrow, it's still better than having jet-lag for a week.

Why you need to eat breakfast at the right time

The reason it is really important to then eat a good breakfast at the right time, is that is the moment when your body clock shifts. Yup, your body clock shifts however many hours in the process of eating that breakfast.  Your hunger clock presses the reset button on your normal circadian body clock.

Now, there is a trick to this: Don't eat breakfast at the time you want to wake up, because you will begin to wake two hours before the time you ate breakfast. I can't remember where I read about this little fact, but once I realised it, suddenly a number of problems I had with the method previously made sense, and I was able to sort them out, largely by not giving in and eating at 05h30 or 06h00 instead of a couple of hours after I wanted to actually wake up.

Basically your body tries to be wide awake at breakfast time, whenever that is.  So if you have breakfast at 06h00 you will wake at 04h00, which was exactly the sort of problem I was having when I first tried this method.  Once I sorted that out, it worked beautifully for me.

How to have the worst jet-lag ever: eat in the middle of the night

Once I realised how that the hunger clock resets the circadian rhythm, and has this 2-hour preamble built into it, the jet-lag horror stories of several friends suddenly made sense: They got up in the middle of the night due to jet-lag, went out and partied, or ate something and watched a movie and then went back to bed.  Every time they did this, they encouraged their body clock to move 2-hours back.  I now believe that their horror stories of week-long jet-lag suddenly made a whole lot more sense. 

How to wake up early every morning if you are "not a morning person"

It also started to occur to me that many people who claim to be "night people" on the basis that they have trouble getting up "early" in fact are poor or late breakfasters.  It might be for many (I don't claim all) that if they force themselves to eat a substantial breakfast within 2 hours of their rising time that they might be able to pull their body clock forward that little bit, and keeping it there.  This is only conjecture, since I am basically a morning person.  I welcome people reporting their experiences.

Honourable Mention at the Global Security Challenge Grand Final

Earlier in the year we entered the Serval Project into the Global Security Challenge.

The theme for this year's challenge was cyber-security, something that we felt was a good fit for the Serval Mesh, with its encrypted-by-default operation, ease of use, and applicability to a wide range of situations.  

The recent revelations of wide-spread wire-tapping by the NSA, GCHQ and others has only served to reinforce the relevance of a secure digital mobile communications platform that never gives unencrypted access to your communications to carriers or other organisations that might be pressured by agencies like the NSA.

So it was very pleasing when we were named finalists last month and invited to pitch at the grand final that was held yesterday here in London.

The grand final itself was an interesting event, with a diverse range of finalists from the UK, the USA, Israel, Sweden, Canada and Spain, and of course ourselves as the sole representative from the Southern hemisphere.  We were also unique in that we were the only social enterprise present.

Each finalist had six minutes to give a pitch, and then a ten minute question and answer session with the judges.  These were combined with the material that we had already submitted to determine the winner of each category.

The day was divided into pre-revenue and post-revenue companies.  I am not sure that they knew quite how to place us, because while we have received over a million dollars in philanthropic funding, we haven't made any commercial sales.

They had to make a decision, and that was to place us in the post-revenue category.  This placed us in the category with a number of innovative start-ups who are already active in the market place, and with business models and solutions that are well aligned with the existing security industry, and representing a formidable field for us to compete against.

So it was very pleasing when it was announced that we had come a close second place in our category. So close in fact, that they took the unusual step of awarding us an Honourable Mention, which apparently is not something that they normally do.  You can get an idea of how unusual this is by the handwritten annotation of honourable mention on the certificate:

While it would have been nice to win outright, it was an amazing affirmation for us to be named in the list of the ten most promising security related endeavours, and then to be recognised in this unusual way.

This is now the third time in the last three months that we have been recognised for our innovation in secure communications, following winning the communications category of The Technology Challenge for Atrocity Prevention, and being the only non-US entity accepted to present at a recent DARPA symposium on secure mesh networks.

Of course, none of this would have been possible without our team, or the support of Flinders University, the Shuttleworth Foundation, the New America Foundation, the NLnet Foundation and the Awesome Foundation and our many individual donors and volunteers.

Over-the-Mesh Updates for Serval Mesh Extender Prototypes

I have just spent the day working on field upgradability for the serval mesh extender prototypes.

Note that the following is how updates for the prototypes work at present. We expect that this may change over time, and is likely to be different in the final products.

Updates to servald will be supported in two different ways.

First, it will be possible to pull the USB memory stick from the unit and copy the new servald binary onto the stick.

Second, if there is no servald binary on the USB memory stick, then it will look for an update via Rhizome.  If there is an update, it will install it.  As with all Rhizome files, the updates are cryptographically protected so that only genuine updates will be accepted.

None of the update functions are automatic, because we want to place updates entirely under user control, as a partial protection against network poisoning attacks.

To effect an update (or to restart the servald on a mesh extender without turning it off and on again), just press the WPS button on the MR3020 router.

Of course, if something stuffs up, for example you have a bad update on the USB memory stick, then hold the WPS button down for 10 seconds and it will "factory reset" back to the original servald that shipped with it.

More as we develop it.

Demoing the Serval Mesh and Mesh Extenders to School Children as part of Australian National Science Week 2013

We were fortunate enough to win a National Science Week grant to run several events in the north of South Australia to demonstrate what we are doing primarily to school students in schools in regional and remote areas.  We were also able to include demonstrating the technology on a steam train.

Our itinerary took us from Adelaide to Port Augusta to Quorn to Hawker over two days.

The idea was to encourage students in the regional centre of Port Augusta and remote towns of Quorn and Hawker to see value and encourage an interest in science, technology, engineering and mathematics (STEM) subjects.

At Port Augusta we worked with 13 to 17 year olds, who quickly worked out how to use the Serval Mesh app on the phones we supplied.  We also had a visit from the local TV station, and we understand that they ran a story the following night.

Several of the guys from Project Horus joined us so that we could put Mesh Extenders up high and communicate with them over longer distances, which we did:

Here are a couple of the switched-on students from Port Augusta Secondary School holding the balloon that Mark and the other Project Horus setup for us.

You can also see our two kids in the foreground.  It was a very welcome change to not have to leave the family behind, but instead share some fun adventures with them.

We then moved along to Quorn and Hawker Area Schools.

Quorn has about 150 students from reception to year 12, and we spoke to all of them in two groups.  All age groups were able to interact with the technology, call each other, and enjoy the balloons organised by Project Horus.  The balloon did pop, though, as the wind picked up to gale force during the morning.

We then moved on to Hawker Area School which has just 47 students.  The wind was way too fast to do any outdoors balloon work, but we still had the students talking and texting within a few minutes in the class room.  We also explained how telephone networks function, even with the littlest kids, explaining how the centralised approach leads to single points of failure.  This was quite fun, with flying sharks making an appearance, as the photo shows:

It is amazing what you can teach six year olds if they don't know that it is material from a second year University course.  We then let the kids draw in their own additions to the diagram, which in addition to the extra flying shark trying to eat the cell tower included other people using mobile and fixed phone lines.  It was particularly impressive to see that a child under the age of 8 had worked out that the fixed phone line needed to be connected to the rest of the network:

By sheer chance Victor Harbor Primary School were staying at Hawker School overnight on their Outback Camp, so we spoke to them as well.

Mark and the guys had setup an indoor balloon for them, and spoke to them about their work, and getting penguins flying close to the edge of space:

Then we all went outside into the howling gale, and released the balloon into the wind to the delight of all. It disappeared so quickly that by the time the camera got organised it was just a small black speck over the tallest tree in this shot:

We rounded out the trip with our final event on the Pichi-Richi Steam Train on the Saturday.  Our idea was to talk to anyone on the train who was willing to listen, and to allow people to talk from end to end on the train.  Problems with the Mesh Extenders meant that we only had end-to-end communications when the train was going around corners, which fortunately is most of the time on the Pichi Richi Pass.

There is very little cellular coverage in the pass, allowing us to capture images like the following showing services available on a mesh-enabled phone where an ordinary phone would have none:

What was particularly nice on the train was one of the students at Quorn Area School was also a car captain on the train, and not only remembered what we had said the previous day, but showed further interest in what we were doing.  So all in all, I'd call the events a success.

Latest iteration of the Mesh Extender

We finally got our hands on some nano-USB memory sticks, and have adjusted the radio placement in the MR3020 routers we are using, so that the lid can go back on without duct-tape, and the whole unit looks much sleeker.

Also, we have some smaller and surprisingly complementary styled rechargeable battery units for them, that even have a percentage charge remaining indication.  All for $17.50 on ebay!

The result is as shown in the following photos.

The battery packs even have a 2-digit LED display to tell you how much battery capacity you have left -- very handy for in the field.

The ability to directly wire the RFD900 into these units simplifies the wiring amazingly, and really is one of the big enablers for being able to assemble the prototypes at an affordable cost:

The biggest outstanding challenge is finding a nano USB memory stick that is both reliable and works sensibly with OpenWRT.  Verbatim ones don't seem to work with OpenWRT properly, while SanDisk ones might still have the "go read-only and die" bug that others have reported.  Time will tell...

ISM 915MHz band availability in different countries

I found this handy table today that lists ISM band availability around 915MHz in a variety of different countries:

http://meshplus.com/?qa_faqs=is-an-fcc-license-required

While closer scrutiny is required, it is the best compilation of this data I have found yet, especially for countries outside of ITU Region 2.